There is a saying in the market that the speed of China's electric vehicle development depends on Musk's disclosure of Tesla's technology.
There is some truth to this statement. In 2014, Musk publicly said: "Our patents belong to everyone", and then opened more than 300 patents on the core technology of electric vehicles to the world. This means that any company can use Tesla's patented technology for free.
According to reliable statistics, car companies, including BYD, Xpeng, BMW, and NIO, are known to use Tesla's patented technology.
And the market volume of electric vehicles has also ushered in explosive growth. In 2014, the global production of about 23.5 million vehicles, of which only 300,000 were electric vehicles, and by 2022, the production of electric vehicles has exceeded 7 million.
Someone once said that Musk's disclosure of patents is a huge loss of wealth, and now it seems that this is to let everyone build momentum for electric vehicles and build momentum for Tesla, and Tesla is still at the forefront of the field of electric vehicles today.
It is no wonder that in the past two days, Musk's visit to China has been able to make waves again.
[Musk's visit to China]
Although he did not participate in the 2024 Beijing Auto Show, on April 28, Musk arrived in Beijing by private jet for an official visit to China. This is also Musk's second visit to China after May last year, and the itinerary was very private, and it was not reported by the media until the plane landed.
Tesla's official document said that the company will continue to cultivate in China, develop together with the industry in artificial intelligence, electric vehicles, energy storage and other fields, and accelerate the implementation of clean energy and autonomous driving technology.
In addition, it was revealed that Musk's trip mainly had the following two things.
First of all, the new entry-level model Model 2 will go on sale in 2025 and will be produced at the Shanghai factory, with a domestic pricing of less than 200,000 yuan.
Secondly, there is the possibility of the implementation of autonomous driving software FSD in China, which we will focus on in this article.
【Tesla FSD】
特斯拉FSD全称为Full Self-Drive，即“完全自动驾驶能力”。
Tesla's self-driving products can be divided into 3 categories, including AP, EAP, and FSD.
Among them, AP is the most basic version, which has been used for free in China;
In order to enhance assisted driving, EAP has added functions such as smart phone summoning, automatic parking, and NOA, and the domestic price is 32,000 yuan;
The highest level of FSD, which was released 4 years ago, includes navigation assistance (NOA), automatic lane change, automatic parking, intelligent summoning, traffic signal recognition, and automatic steering of urban roads (based on navigation routes), which are currently only available in North America, and the main reason why they have not been promoted in China is that it involves more obvious data security issues.
Just before Musk's arrival in Beijing, the China Automobile Association had issued the "Intelligence on the Testing of 4 Safety Requirements for Automotive Data Processing", which clearly pointed out that all Tesla models fully meet the compliance requirements, and it is also the only foreign-funded car company that meets the compliance requirements, and other car companies include BYD, Ideal, Lotus, Hezhong, and Weilai.
These four requirements include: anonymization of face information outside the car, no collection of cockpit data by default, in-vehicle processing of cockpit data, and significant notification of personal information processing.
In fact, Tesla established a data center in Shanghai as early as 2021 to achieve localized data storage.
In the field of intelligent driving, Tesla is still the first echelon.
【Tesla's Intelligent Driving Authority】
Musk believes that humans can drive cars with their eyes and brains, so equipping cars with vision devices and computing systems can achieve autonomous driving.
So Tesla FSD is a pure vision solution, not a lidar. Logically, the vehicle collects video data through the camera, and relies on computing chips and leading algorithms to process the data to generate the vehicle's control instructions, including steering angle, braking force and acceleration. The advantage is that the cost is low, but the identification is limited and greatly affected by the weather.
In China, the mainstream is the lidar solution. Taking Huawei as an example, there are already up to 3 lidar intelligent driving solutions, which have the advantage of breaking through visual limitations, which is also in line with the complex driving environment in China, but the cost is high.
Tesla's FSD has the advantage of full-stack self-development, including self-developed chips, powerful data acquisition capabilities, and huge computing power, Tesla's 2022 training platform computing power scale is 2 EFLOPS, and it is planned to reach 100 EFLOPS by October this year. For reference, the computing power of the ADS2.0 training platform disclosed at the M7 press conference in September 2023 is 1.8 EFLOPS.
Throughout the history of FSD, Tesla began to develop self-driving systems in 2013;
In 2014, HW1.0 was launched;
In 2019, it was equipped with a self-developed FSD (full autonomous driving) chip for the first time, bringing stronger computing power and launching HW3.0;
In 2020, Tesla developed a new training computer, Dojo, and introduced the BEV+Transformer architecture, marking the entry of Tesla's self-driving technology into the era of large models.
In 2021, Tesla officially released the beta version of FSD;
On March 31, 2024, bidding farewell to the three-and-a-half-year beta phase, Tesla pushed the FSD V12 (Supervised) version to some users in the United States.
[Intelligent driving still needs to be invested]
Of course, Tesla's FSD is not a perfect smart drive. Just before Musk's trip to China, the U.S. auto safety regulator announced on April 26 that it had launched an investigation into Tesla's models equipped with Autopilot's assisted driving system, because there were still multiple collisions after Tesla recalled 2 million vehicles and upgraded the Autopilot system in December last year. A few days ago, Musk released a new tweet, saying that Tesla will invest about $10 billion in comprehensive training and inference artificial intelligence this year, the latter mainly for cars.
Obviously, autonomous driving still requires a lot of money and time investment, but it is also a technology that will definitely break through, and Tesla, as the first echelon, is naturally attracting attention. As for domestic manufacturers to distance themselves from Tesla, they must find their own path in the competition.
On April 23, Huawei's Yu Chengdong said in a high-profile manner that the Zhijie S7 team went to the United States to compare the intelligent driving system of Zhijie S7 with Tesla's FSD, "After testing, we are confident that the intelligent driving system of Zhijie S7 is still the best in the world."
In addition, after Musk came to China, He Xiaopeng, chairman of Xpeng Motors, also mentioned that he "welcomed" FSD into China.
So, are you optimistic about Tesla's FSD? How long do you think it will take for us to drive a "fully self-driving" car?